Skip to main content

Programme

Competitions

Two accepted competitions challenge participants to develop and evaluate autonomous trading systems in realistic financial environments.

ICAIF '26 Programme

Accepted Competitions

Competition websites, participation dates, and registration instructions will be added as they become available.

Live trading competition Testnet environment

Perpetual Alpha: An On-Chain Algorithmic Trading Competition on the Hyperliquid Testnet

Perpetual Alpha is a live algorithmic-trading competition in which teams design, implement, and operate autonomous trading agents on the Hyperliquid testnet. Organizers will deploy dedicated HIP-3 perpetual futures markets covering major crypto assets and European equity underlyings.

Teams compete over multiple weeks under identical starting conditions using non-monetary testnet capital. The final ranking combines realized profit and loss (40%), the strategy's risk profile measured through return volatility (30%), and algorithmic sophistication assessed by an expert jury (30%). Winners will be announced at ICAIF '26, where top teams will present their approaches.

Organizers

  • Andrea Prampolini

    Intesa Sanpaolo

  • Edoardo Vittori

    Intesa Sanpaolo

  • Manuel Naviglio

    Scuola Normale Superiore

  • Francesco Tarantelli

    Università di Bologna

Competition format

Environment Hyperliquid testnet · live on-chain perpetual futures trading

Participation details to be announced
Live benchmark Portfolio management

ACM ICAIF 2026 Trading Agent Competition

This live portfolio-management benchmark welcomes autonomous trading agents of any design. Participants manage a long-only portfolio across 30 U.S.-listed equities in six sectors, using standardized market and fundamental features and, optionally, public financial news. Agents submit target portfolio weights for execution at the next market open.

Every agent starts with USD 1,000,000 in cash and operates under common constraints: a 30% cap per asset, gross exposure of no more than 100%, and a 0.1% transaction fee on purchases and sales. Participants receive 2020–2024 training data, an environment SDK, baseline agents, and validation on held-out 2025 data before final evaluation on newly released live market data.

Performance is assessed across nine measures covering profitability, risk-adjusted returns, downside risk, and execution quality: cumulative return, daily win rate, Sharpe ratio, Sortino ratio, maximum drawdown, value at risk, expected shortfall, turnover, and violation rate.

Organizers

  • Xinyu Xi

    National University of Singapore

  • Yifan Bao

    National University of Singapore

  • Ha Cong Nga

    National University of Singapore

  • Qiang Wang

    National University of Singapore

  • Yihao Ang

    National University of Singapore

  • Anthony K. H. Tung

    National University of Singapore

  • Yueju Han

    South China University of Technology

  • Xin Zhang

    South China University of Technology

  • Hao Ni

    University College London

  • Lukasz Szpruch

    University of Edinburgh

Competition format

Environment Shared SDK · held-out validation · live market-data evaluation

Participation details to be announced